A high-precision and low-latency loading method for aerodynamic load in model experiment

Abstract

The application discloses a kind of model experiment aerodynamic load high-precision and low-delay loading method, belong to the field of ocean engineering, comprising: obtaining the static characteristic curve and dynamic characteristic parameter of actuator;According to target working condition, generate target aerodynamic thrust sequence, based on static characteristic curve and dynamic characteristic parameter, calculate offline inverse control sequence by inverse dynamics model;In real-time loading stage, according to current physical time and pure lag time, determine look-ahead time, based on look-ahead time, carry out time domain interpolation in offline inverse control sequence to obtain the control instruction of current time and send to actuator.The application combines the hybrid control strategy of offline inverse model compensation and online time domain interpolation look-ahead, effectively eliminates the phase lag and amplitude attenuation caused by physical inertia of actuator, realizes high-precision, low-delay aerodynamic load loading, and is suitable for wind load equivalent simulation in scale model test of floating offshore wind turbine.

Description

Technical Field

[0001] This invention belongs to the field of marine engineering, and in particular relates to a high-precision and low-delay loading method for aerodynamic loads in model experiments. Background Technology

[0002] As offshore wind power expands into deeper waters, floating wind turbines are becoming the mainstream trend. In the development of floating wind turbines, scaled-down model experiments in wave tanks are a crucial step in verifying their dynamic performance. However, scaled-down experiments suffer from the well-known "Froude-Reynolds number scaling paradox": to satisfy hydrodynamic similarity, the model must adhere to the Froude similarity criterion, but directly scaling down the physical blades at this ratio leads to a sharp decrease in the Reynolds number and a severe deficiency in aerodynamic thrust. To address this issue, the current mainstream international approach employs "hybrid model experiments" or "software-in-the-loop" techniques. These involve installing aerodynamic equivalent loading devices such as ducted fans or servo motor propellers on the top of the model tower. The control system calculates the target aerodynamic thrust in real time and drives the fan to generate thrust, thus replacing the aerodynamic load on the actual physical blades.

[0003] Existing hybrid model testing techniques still face significant bottlenecks in achieving high-precision, low-latency aerodynamic load loading. On the one hand, the actuators and their drive motors have considerable physical inertia. When simulating high-frequency turbulent winds (such as the Kaimal spectrum) or processing high-frequency motion of the platform, the response speed cannot keep up with the changes in the target command, resulting in significant phase lag in the thrust and even non-physical system instability. On the other hand, existing compensation algorithms, such as Kalman filtering and Smith predictors, suffer from high complexity, reliance on accurate time-delay models, or poor robustness, making them difficult to implement efficiently on low-cost industrial controllers. Furthermore, existing systems rely on discrete data exchange with fixed time steps, making it difficult to perfectly synchronize the control cycle with the simulation step size. The traditional "lookup table method" ignores the continuity between time steps, resulting in a stepped output signal, introducing additional quantization noise, and reducing loading smoothness. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a high-precision and low-delay aerodynamic load loading method for model experiments, comprising: Obtain the static characteristic curve and dynamic characteristic parameters of the actuator, wherein the dynamic characteristic parameters include time constant, pure time delay, and amplitude gain correction coefficient; A target aerodynamic thrust sequence is generated based on the target operating conditions, and an offline inverse control sequence is calculated using an inverse dynamics model based on the static characteristic curve and the dynamic characteristic parameters. During the real-time loading phase, the look-ahead time is determined based on the current physical time and the pure time delay, and time-domain interpolation is performed on the offline inverse control sequence based on the look-ahead time to obtain the control command at the current moment. The control command is sent to the actuator to drive the actuator to generate the actual thrust corresponding to the target aerodynamic thrust sequence.

[0005] Optionally, obtaining the static characteristic curve of the actuator includes: Send a PWM step signal covering the entire range to the actuator and record the steady-state thrust value; The mapping relationship between the PWM value and the steady-state thrust value is fitted using a double Logistic superposition function to obtain the static characteristic curve parameters.

[0006] Optionally, obtaining the dynamic characteristic parameters includes: Generate the target wind spectrum fragment corresponding to the target working condition as a pre-calibration sequence; The pre-calibrated sequence is sent to the actuator, and the measured thrust response is collected simultaneously. Based on the pre-calibrated sequence and the measured thrust response, the system is identified and fitted to obtain a first-order plus pure time-delay model. The time constant is extracted from the transfer function of the first-order pure time delay model, and the pure time delay is calculated through the cross-correlation function. The amplitude gain correction coefficient is calculated through the power spectral density energy ratio.

[0007] Optionally, the system identification based on the pre-calibration sequence and the measured thrust response includes: Import the PWM input sequence and the measured thrust response sequence into the system identification tool; The input and output data are fitted into a series model of a first-order inertial element and a pure time delay element using the least squares method.

[0008] Optionally, based on the static characteristic curve and the dynamic characteristic parameters, an offline inverse control sequence is calculated using an inverse dynamics model, including: Based on the target aerodynamic thrust sequence and the static characteristic curve, the reference PWM sequence is obtained by inverse solving. A dynamic compensation term is calculated based on the time constant to counteract the physical inertia of the actuator; The dynamic compensation term is corrected according to the amplitude gain correction coefficient; The corrected dynamic compensation term is superimposed on the reference PWM sequence to generate an offline inverse control sequence.

[0009] Optionally, the dynamic compensation term is calculated based on the time constant, including: Calculate the first derivative of the target aerodynamic thrust sequence and multiply it by the time constant to obtain the dynamic lead used to compensate for the inertial lag of the actuator.

[0010] Optionally, time-domain interpolation is performed on the offline inverse control sequence based on the look-ahead time, including: Obtain the current physical time provided by the high-precision timer and add the pure time delay to obtain the look-ahead time; Based on the look-ahead time, locate adjacent discrete data points in the offline inverse control sequence; Linear interpolation is performed on the adjacent discrete data points to obtain the floating-point PWM instruction value corresponding to the look-ahead time.

[0011] Optionally, it also includes: Real-time acquisition of actual thrust data generated by the actuator; Based on the deviation between the actual thrust data and the target aerodynamic thrust sequence, determine whether the current operating condition has changed significantly; When it is determined that the operating conditions have changed significantly, the steps of obtaining the static characteristic curve and dynamic characteristic parameters of the actuator are repeated to update the dynamic characteristic parameters.

[0012] On the other hand, the present invention also provides an electronic device including a memory, a processor, and a computing program stored in the memory and executable on the processor, wherein the processor implements the method when executing the computing program.

[0013] On the other hand, the present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method.

[0014] Compared with the prior art, the present invention has the following advantages and technical effects: This invention obtains specific dynamic characteristic parameters of the actuator by identifying in-situ parameters based on operating conditions. Combined with offline inverse dynamics model calculation and online time-domain interpolation look-ahead compensation, it effectively overcomes the phase lag problem caused by the physical inertia of the actuator and achieves zero-phase delay tracking of the target aerodynamic thrust. By eliminating timing jitter and quantization noise caused by discretized control through continuous time-domain linear interpolation, the smoothness of control is significantly improved. At the same time, the use of dual logistic static characteristic fitting and amplitude gain correction ensures high-precision reproduction of thrust amplitude under complex wind spectra, greatly reducing the mean absolute error, root mean square error and maximum absolute error, providing a high-fidelity aerodynamic load simulation environment for floating wind turbine model tests. Attached Figure Description

[0015] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart illustrating the overall implementation of the algorithm in this embodiment of the invention. Figure 2 This is a schematic diagram of the inverse dynamics model compensation according to an embodiment of the present invention; Figure 3 This is a diagram illustrating the online time-domain interpolation and hysteresis compensation mechanism in an embodiment of the present invention. Figure 4 This is a diagram of the dual Logistic static characteristic curve of the motor according to an embodiment of the present invention; Figure 5 This is a comparison chart of the temporal tracking effects of embodiments of the present invention; Figure 6 This is a tracking error diagram according to an embodiment of the present invention. Detailed Implementation

[0016] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0017] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0018] Example 1 This embodiment provides a high-precision and low-delay aerodynamic load loading method for model experiments, including: Figure 1 As shown, the stages are: system modeling and parameter identification, offline inverse model sequence generation, and online real-time experimentation.

[0019] Obtain the static characteristic curve and dynamic characteristic parameters of the actuator, wherein the dynamic characteristic parameters include time constant, pure time delay, and amplitude gain correction coefficient; A target aerodynamic thrust sequence is generated based on the target operating conditions, and an offline inverse control sequence is calculated using an inverse dynamics model based on the static characteristic curve and the dynamic characteristic parameters. During the real-time loading phase, the look-ahead time is determined based on the current physical time and the pure time delay, and time-domain interpolation is performed on the offline inverse control sequence based on the look-ahead time to obtain the control command at the current moment. The control command is sent to the actuator to drive the actuator to generate the actual thrust corresponding to the target aerodynamic thrust sequence.

[0020] Specifically: The process of modeling and parameter identification of the operating system includes: To address the issues of poor universality and low accuracy of parameters obtained by traditional step response calibration methods under complex wind spectra, this embodiment employs an "in-situ identification method based on target operating conditions." This method directly utilizes the target wind spectrum segment as the excitation signal to obtain the system parameters that best fit the current experimental conditions.

[0021] Static characteristic parameter calibration: The computer-controlled motor sequentially outputs PWM step signals covering the entire range. After the thrust stabilizes, the steady-state thrust value is recorded. The "PWM-thrust" data is fitted using a double Logistic superposition function (e.g., ...). Figure 4 ), obtain static mapping parameters ( This is used for subsequent static inverse kinematics calculations. The fitting formula is as follows: ; In-situ identification of dynamic characteristic parameters (obtaining τ, Delay, Gain): A Kaimal target wind spectrum segment of duration τ is generated as a "pre-calibration sequence". This sequence contains the true frequency components and amplitude variation characteristics of the target operating condition. The target sequence is sent to the motor system, and the force sensor synchronously records the input command sequence PWM and the measured thrust response. The collected PWM values ​​and Import the data into MATLAB for system identification analysis. Use the system identification toolbox to fit the input and output data into a first-order plus pure time delay model. .

[0022] Parameter extraction: The time constant T is directly extracted from the denominator of the fitted transfer function. This method yields τ, which contains the average dynamic response characteristics of the motor at a specific frequency of the Kaimal spectrum, superior to the value obtained from a single step signal. By calculating PWM and... Find the lag time L corresponding to the maximum value of the cross-correlation function; compare the power spectral density (PSD) of the measured response with the PSD of the target spectrum, calculate the energy ratio between the two, and determine the amplitude gain correction coefficient to ensure energy consistency in the dynamic process.

[0023] like Figure 2 As shown, the process of generating the offline inverse dynamics control sequence includes: This stage utilizes the specific parameters for the operating conditions identified in the first stage to perform calculations, ensuring the accuracy of the control sequence.

[0024] 1. Generate target wind spectrum: Generate a complete kaimal target thrust sequence based on the experimental scaling ratio. .

[0025] 2. Inverse Model Core Calculation: The baseline PWM is calculated using a static double Logistic model; based on the identified operating condition-specific time constant τ, the dynamic compensation term is calculated. The identified gain coefficient Gain is introduced to correct the dynamic compensation term. This step effectively eliminates overshoot or undershoot caused by model parameter mismatch.

[0026] 3. Synthesize the final sequence: Generate the final offline inverse control PWM sequence.

[0027] like Figure 3 As shown, the online high-precision time-shift interpolation loading process includes: This stage applies the offline computed sequence to the hardware and uses the identified lag parameters for time axis compensation.

[0028] 1. Experiment Start-up and Timing: Start the high-precision timer to acquire physical time. .

[0029] 2. Look-ahead compensation: Calculate the look-ahead time using the identified precise delay time (Delay). .

[0030] 3. Linear interpolation: For offline sequences... Perform linear interpolation to obtain smooth control commands.

[0031] 4. Closed-loop verification: Data is recorded in real time. If a drastic change in operating conditions is detected, the "in-situ identification" process of the first stage can be re-executed to update the parameters. The control effect is shown in Table 1.

[0032] Table 1 5. Experimental Results and Analysis like Figure 4 The figure shows the calibration results of the static characteristic curve of the actuator. The horizontal axis represents the input PWM pulse width (in microseconds), and the vertical axis represents the generated steady-state thrust (in Newtons). Ten experiments were conducted, and the average value of these ten experiments was taken as the final fitting result. As can be seen from the figure, the input-output relationship of the actuator exhibits obvious nonlinear characteristics, and a dead zone exists in the low PWM range. This invention uses a double logistic superposition function to fit the measured data points. As shown by the red fitting curve in the figure, the fitting curve highly overlaps with the black measured data points, and the adjusted R-squared reaches 0.99965, indicating that the model can extremely accurately describe the nonlinear static gain of the actuator, providing a high-precision benchmark for subsequent inverse model calculations.

[0033] like Figure 5The figure shows a comparison of time-domain thrust response under Kaimal turbulent wind conditions. The solid black line represents the target aerodynamic thrust sequence, the dashed blue line represents the response curve using the traditional direct control method, and the solid red line represents the response curve using the method of this invention. Observing the magnified details reveals that the dashed blue line lags significantly behind the target curve at the peaks and troughs, exhibiting a physical delay of approximately 200-300 ms, failing to respond promptly to high-frequency turbulent changes. In contrast, the solid red line almost completely overlaps with the black target curve, indicating that time-domain interpolation with look-ahead time effectively compensates for the lag time of the physical system, achieving "zero-phase" tracking.

[0034] like Figure 6 As shown, the instantaneous tracking error (i.e., measured value minus target value) under two control strategies is intuitively displayed. The blue dashed line represents the error of direct control, which fluctuates drastically, with a maximum error exceeding 0.6N, and the error polarity frequently reverses, indicating that the system has serious overshoot and lag. The red solid line represents the control error of the present invention, which is always limited to a very small range (within ±0.40N) and converges rapidly under sudden operating conditions.

[0035] In this embodiment, the PWM value to be sent is calculated and sent to the Arduino Mega 2560 core control board via serial port. The control board sends the PWM signal to the electronic speed controller, which controls the motor to generate thrust. The force sensor reads the thrust data and feeds it back to the host computer to verify the tracking effect.

[0036] Example 2 This embodiment provides a high-precision and low-delay aerodynamic load loading method for model experiments, including: 1. Floating offshore wind turbine (FOWT) wind-wave coupled dynamic response tank model test: In the research and development verification of deep-sea floating wind turbines, accurately evaluating the motion response of the floating platform under the combined effects of wind and waves is crucial. The control algorithm of this invention is integrated into the host computer controller of a "software-in-the-loop" hybrid model test system. Before the test, the "in-situ identification based on target operating conditions" function of this invention is used to allow the actuator to pre-run a Kaimal wind spectrum, automatically identifying and locking the physical inertial parameters (time constant τ) and communication lag time (Delay) of the current wind turbine system. In the formal test, the algorithm uses an "offline inverse model + online time-domain interpolation" strategy to calculate in real time and send PWM commands with look-ahead compensation to the motor. This effectively eliminates the thrust lag phenomenon common in traditional PID control, ensuring that the aerodynamic thrust and wave load on the model are strictly matched in time, accurately reproducing the unique "aerodynamic damping" effect of the floating wind turbine, thereby obtaining high-precision pitch and heave motion response data.

[0037] 2. Dynamic wind load simulation for offshore platforms and large surface structures: For FPSOs (Floating Production Storage and Offloading) or semi-submersible drilling platforms, the superstructure has a large wind-exposed area, and transient wind loads under gusts or squall line conditions pose a significant challenge to the safety of the mooring system. Such tests typically employ multi-fan arrays to simulate large-area wind fields. This algorithm can serve as the core control logic, controlling each fan unit in the array individually. To address potential response differences between different fan units, the algorithm introduces an independent "amplitude gain correction coefficient" to ensure that the output thrust of each channel can track high-frequency abrupt changes without overshoot. When simulating extreme gust conditions, this algorithm achieves millisecond-level thrust response, avoiding load amplitude attenuation or overshoot caused by actuator inertia. This provides a more rigorous and accurate testing environment for evaluating the maximum mooring tension and structural transient response of offshore platforms under extreme sea conditions, effectively solving the problem that traditional static lookup table methods cannot realistically simulate dynamic gusts.

[0038] The beneficial effects of this embodiment are: 1. Extremely low phase delay: achieved through delay parameters With the introduction of this algorithm, the algorithm is essentially sending "future" instructions to the actuator. When the instructions are transmitted and translated into mechanical actions, physical time has just elapsed. This ensures that the generated force is perfectly synchronized with the target force at the current moment.

[0039] 2. High-precision amplitude-frequency characteristic reproduction: Through the dynamic compensation term and gain correction coefficient in the inverse model, the power spectral density (PSD) of the measured force is highly consistent with the theoretical Kaimal spectrum, and not only is the average value accurate, but the energy distribution of the pulsation characteristics is also accurate.

[0040] 3. High control smoothness: By introducing a linear interpolation algorithm, even if the control loop experiences slight jitter due to Windows system scheduling or control frequency adjustment, the output signal remains smooth and continuous, avoiding control noise caused by discretization.

[0041] 4. Strong system adaptability: The key parameters in the code (lag time, gain correction, time constant) are all defined as constants, which can be quickly adjusted and calibrated according to wind turbine models of different sizes or different wind tunnel environments without modifying the core algorithm structure.

[0042] 5. High Parameter Identification Accuracy and Strong Adaptability to Operating Conditions: This invention abandons the traditional parameter calibration method based on a single step signal and innovatively adopts an "in-situ identification based on target operating conditions" strategy. By running a target spectrum segment for system identification before the formal experiment, the system parameters under that specific frequency and load characteristics are directly obtained. This method effectively solves the problem of parameter failure caused by nonlinear changes in the dynamic characteristics of the motor system under different wind spectrum operating conditions, ensuring the optimality of the algorithm parameters under the current experimental conditions.

[0043] On the other hand, this embodiment also provides an electronic device, including a memory, a processor, and a computing program stored in the memory and executable on the processor, wherein the processor implements the method when executing the computing program.

[0044] On the other hand, this embodiment also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method.

[0045] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A high-precision and low-delay aerodynamic load loading method for model experiments, characterized in that, include: Obtain the static characteristic curve and dynamic characteristic parameters of the actuator, wherein the dynamic characteristic parameters include time constant, pure time delay, and amplitude gain correction coefficient; A target aerodynamic thrust sequence is generated based on the target operating conditions, and an offline inverse control sequence is calculated using an inverse dynamics model based on the static characteristic curve and the dynamic characteristic parameters. During the real-time loading phase, the look-ahead time is determined based on the current physical time and the pure time delay, and time-domain interpolation is performed on the offline inverse control sequence based on the look-ahead time to obtain the control command at the current moment. The control command is sent to the actuator to drive the actuator to generate the actual thrust corresponding to the target aerodynamic thrust sequence.

2. The method according to claim 1, characterized in that, The process of obtaining the static characteristic curve of the actuator includes: Send a PWM step signal covering the entire range to the actuator and record the steady-state thrust value; The mapping relationship between the PWM value and the steady-state thrust value is fitted using a double Logistic superposition function to obtain the static characteristic curve parameters.

3. The method according to claim 1, characterized in that, The acquisition of dynamic characteristic parameters includes: Generate the target wind spectrum fragment corresponding to the target working condition as a pre-calibration sequence; The pre-calibrated sequence is sent to the actuator, and the measured thrust response is collected simultaneously. Based on the pre-calibrated sequence and the measured thrust response, the system is identified and fitted to obtain a first-order plus pure time-delay model. The time constant is extracted from the transfer function of the first-order pure time delay model, and the pure time delay is calculated through the cross-correlation function. The amplitude gain correction coefficient is calculated through the power spectral density energy ratio.

4. The method according to claim 3, characterized in that, The system identification based on the pre-calibrated sequence and the measured thrust response includes: Import the PWM input sequence and the measured thrust response sequence into the system identification tool; The input and output data are fitted into a series model of a first-order inertial element and a pure time delay element using the least squares method.

5. The method according to claim 1, characterized in that, Based on the static characteristic curve and the dynamic characteristic parameters, an offline inverse control sequence is calculated using an inverse dynamics model, including: Based on the target aerodynamic thrust sequence and the static characteristic curve, the reference PWM sequence is obtained by inverse solving. A dynamic compensation term is calculated based on the time constant to counteract the physical inertia of the actuator; The dynamic compensation term is corrected according to the amplitude gain correction coefficient; The corrected dynamic compensation term is superimposed on the reference PWM sequence to generate an offline inverse control sequence.

6. The method according to claim 5, characterized in that, The dynamic compensation term is calculated based on the time constant, including: Calculate the first derivative of the target aerodynamic thrust sequence and multiply it by the time constant to obtain the dynamic lead used to compensate for the inertial lag of the actuator.

7. The method according to claim 1, characterized in that, Based on the look-ahead time, time-domain interpolation is performed on the offline inverse control sequence, including: Obtain the current physical time provided by the high-precision timer and add the pure time delay to obtain the look-ahead time; Based on the look-ahead time, locate adjacent discrete data points in the offline inverse control sequence; Linear interpolation is performed on the adjacent discrete data points to obtain the floating-point PWM instruction value corresponding to the look-ahead time.

8. The method according to claim 1, characterized in that, Also includes: Real-time acquisition of actual thrust data generated by the actuator; Based on the deviation between the actual thrust data and the target aerodynamic thrust sequence, determine whether the current operating condition has changed significantly; When a significant change in operating conditions is determined, the steps of obtaining the static characteristic curve and dynamic characteristic parameters of the actuator are repeated to update the dynamic characteristic parameters.

9. An electronic device comprising a memory, a processor, and a computing program stored in the memory and executable on the processor, characterized in that, When the processor executes the computing program, it implements the method of any one of claims 1-8.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1-8.

Images